Learnlab
Welcome to LearnLab!
This web service facilitates the extraction and generation of graph-based representations from PDF-format textbooks. We have written a special parsing algorithm for the system to effectively capture and model the hierarchical structure of textbook content: pages, lines, sentences, and identified concepts. These elements are treated as discrete nodes connected by relational edges (Bloom Edges) within the constructed graph. Future enhancements will extend support to more complex structural elements, including sections, chapters, and paragraphs. The resulting graph data is accessible for visualization via the web interface and can be exported to Excel for further analysis.
Each session is associated with a “Task ID”. You can use it to retrieve submitted tasks that have already produced results or to check the status of a task that is still being processed.
To submit a task, you can do the following steps:
  • From the "Select Source" menu, you have two options to select a textbook:
    • You can upload your own PDF textbook file using “Upload a Textbook (PDF),” or
    • You can select an existing textbook using “Select a Preloaded Textbook.”
  • If you choose to upload your own PDF textbook file, any available metadata in the selected file will be extracted. You can manually provide the missing data if you like.
  • If you choose a preloaded textbook, the textbook’s metadata will be provided.
  • In the “Pages to process” textbox, you can select the pages to process (see the examples below).
  • From the “Pre-Processing Lists” box, we provide a filtering mechanism through two types of lists:
    • Substitution List: this list must contain n-gram concepts (i.e., concepts which consist of multiple words). The assumption here is that the provided concepts have already been identified and included in the textbooks' targeted domain. They will be used to enhance identification and reduce future manual filtering. A default list is provided for the Computer Science domain (3700+ concepts, reviewed manually).
    • Ignore List: this list contains concepts to be filtered out. It functions like a typical stopword list. A default list is provided (1000+ concepts – collected manually).
  • Click the “Submit” button.
  • Once a task is submitted, you will receive status updates for each processing stage.
Textbook

Task States : S Success R In Progress P Pending F Failed

Bloom Lab

You can perform two tasks through two options:
1. Bloom Associations: You can add, edit or upload Bloom Associations. These associations are essentially Bloom links between existing concept and question nodes. Source nodes can be of type [Concept] or [Question] while target nodes must be only of type [Concept].
2. Questions: You can add, edit or upload Questions that would target concepts.
Please note that ALL Bloom Associations, whether between concepts or between questions and concepts, should be entered through the "Bloom Association" option. The "Questions" option is only for adding [Question] nodes.
The required columns in the [.csv] formatted file are:
10%
malformed entries (e.g. missing values) and nonexistent source nodes or target concepts will be ignored.
[.csv] file headers' exact names as shown above are required.
Enter the {Source Node [Concept Name, Question Name]} and {Target Concept}, and select their respective Bloom Levels:
Question Name
Question Text
{Question Name} must be unique.
** Please note: the above visualization is currently limited to 50 nodes selected sequentially (hence the presence floating nodes, sometimes) and presented for demonstration purposes at this point of time.
Status |
Ok
* Longer processing time

# Excel File Description Generation Date Status
No Excel files have been generated yet.